Library
U S. Naval Postfrraduate SchooT
Monterej, Califoniia '^
AK INVKSTIQATION OF THE iSFFBCTS 0/ TUHBULENT
qjJSNCHING IN A CM-TIPE COMBUSTION CHAMBER
A Thesis
Submitted to the Graduate Faculty
of the University of Minnesota
by
Robert S. Hutches
LT, U. S. Navy
In Partial Fulfillment of the Requirements
for the Degree of
Master of Science in Aeronautical Ingineering
May 1954
Thesis
Lib-
IVIoiiteie}-, California ^^
ACKNOWLEDGMENTS
The author wishes to express his appreciation to
Dr. Newman A. Hall, Professor Thomas B. Murphy, and Mr.
Boward McManus for their Interest and advice; to LCDR
Winston L. Miller and LT Robert J. Barnes, USN, for their
assistance in operating the equipment and reducing data;
and to the U. S. Naval Postgraduate School for sponsoring
the attendance of the author at the University of Minnesota
during this period of graduate study.
25006
lii
TABLE OF CONTENTS
Page
Symbols and Notation iv
Summary 1
Introduction 2
Equipment 6
Instrumentation 7
Experimental Technique 9
Estimated Errors ..... 13
Results 16
Bibliography 20
Figures 21
Appendices:
"%. Fuel Specifications 43
B. Efficiency Calculations 46
C. Pressure Loss Computations 46
iT
SYMBOLS
A - Area, ft2.
a - Acoustic velocityi ft/eec.
Cp - Specific heat at constant pressure, BTU/l'b.^R.
D - Diameter, feet.
g - 32.2 ft/aec^.
k - Constant.
n - Maes flow rate, lb/sec.
M - Mach number.
P - Pressure.
<il, - Lower heating value of fuel, BTU/lb.
E - Gas constant, = 53.34 ft/oR.
T - Absolute ten^erature, °H.
y - Velocity, ft /sec.
z - Distance from most forward part of the combustion chamber
to the centerline of the first quench air port.
^ - Batio of Cp/C^.
ilk - Increment or change.
Af^j- Pressure drop across metering orifice number 2(3).
J- - Density, lb/ft"''.
. - Efficiency.
Subscripts
( )]^- Ambient air.
()2- Air upstream of main air suprly metering orifice.
2 - Air upstream of vaporizer tube air supply metering orifice.
- Air.
a
g - Exit Cross-Section.
J. - Fuel.
^ - Stream tube in exit cross-section.
uj - Mixture or combustion products.
g - Static or Stream values.
^ - Total or Stagnation values.
th~ Theoretical.
Superscripts
()l,2,etc. _ R^f.gj. ^^ numbered references in BIBLIOGfiAPHT.
SUWMAHY
Secondary air injection orifices in an experimental can-
type combustion chamber usinf^ a vaporizer tube for fuel injection
are modified to produce turbulent mixing of the excees air with
the combustion products. The results of this investigation are
conpared with those of a previous project and the following
results are noted:
1. Burner thermal efficiences are very slightly reduced.
2. Mixing is thorough, as determined by flame patterns
and temperature data.
3. Even ten^erature and velocity profiles are obtained
at the exit cross-section of the combustion chamber.
4. This type of quenching would permit shortening of
a 20-inch combustion chamber "by at least 2-inche8,
or 10^.
5* Total pressure losses were doubled, increasing from
an average value of 5.6^ to 11.3^ of the inlet total
pressure.
INTKODUCTIOM
The trend toward th^ replaceraent of the piston-engine "by
turto-Jet and turbo-prop power plants in all phases of high-speed
and long-range aircraft has placed renewed emphasis on the investi-
gation and development of the constant -pressure, continuous-flow
combustion chamber. Weight and space limitations on aircraft com-
ponents are equally applicable to the jet engine and its con5)onent
parts. Operating requirements for a gas turbine power plant de-
mand smooth, dependable long-life operation over a wide range of
altitude, engine and aircraft speeds, and periods of acceleration
and deceleration. Further design requirements are high heat re-
lease per unit volume of combustion chamber, with high combustion
efficiency and minimum pressure loss.
CTonslderablo investigation has been done on the chemistry
of combustion, propagation of flame fronts, and associated phe-
nomena. Unfortunately this work is not completely applicable to
the type of combustion which occurs in the combustion-chamber of
a gas- turbine power plant. Here, the flame front actually con-
sists of numerous small individual flame fronts. This situation
does not lend itself to easy analysis. Therefore, construction
and final configuration of combustion chambers in gas-turbines
have been largely a matter of experimentation and testing.
Mass and heat transfers in the combustion process involve
four steps: formation of the combustible mixture, ignition or
3
start of combustion, flame movement or propagation of combustion,
and final mixing of the comt)U8tion products with excess air.^
Some factors affecting the formation of a combustible
mixture are here listed. The Bureau of Standards has determined
that combustion efficiency of various fuels in moving air increases
with an increase in fuel volatility. The rate of evaporation of a
droplet of volatile fuel is proportional to the vapor pressure of
the fuel, the absolute temnerature, and the air turbulence, and is
inversely proportional to the molecular weight of the fuel,*" The
air requirements for combustion are proportional to the molecular
weight of the fuel, and the time required to form a combustible
mixture of air and fuel vapor is directly proportional to the fuel
droplet size and is inversely proportional to the relative velocity
4.
between j^he droplet and the air.
L^ition and propage,tion of combustion might be considered
together as a chemical reaction between air and fuel vapor. The
cofTibustion process consists of the breaking-down of the complex
fuel hydrocarbons into lower molecular weight oxides. During these
chain reactions, chain carriers are formed and heat is liberated
to further the reaction until combustion is complete. While the
rate of these reactions is dependent on many factors, it has been
determined that reaction rates are proportional to the absolute
temperatures at which they occur and are inversely proportional
to the molecular weight of the fuel."^*-
4
Prom the above statemente it is seen that a vaporizer tuhe
operating with a low molecular weight, highly volatile fuel in
turbulent air should be an optimum method of introducing fuel into
a combustion chamber. Barnes and Miller^ have conducted investi-
gations using a vaporizer tube in an experimental combustion
chamber, and report high thermal efficiencies as compared to a
spray nozzle type fuel injection system. '^
The final mixing of the combustion products with excess
air occurs in the so-called secondary air zone, the airflow having
been divided into two main portions. The primary air passes
through the primary zone, encoii|}a88ing steps 1, 2, and 3 above.
The secondary air by-passes the primary zone and is injected down-
stream, cooling the products of combustion to obtain a combustion
chamber «xlt temperature profile in which ten^jerature should not
vary more, than 5 per cent from Its average value, and in which the
maximum temperature does not exceed approximately 1700°^.
This secondary air must be injected at hi^ velocities,
of the order of 200-300 feet per second, as compared with 15-50
feet per second in the primary zone. The high velocity is needed
to obtain penetration of the secondary air into the primary com-
bustion products in a very short period of time, since the space
and weight limitations on a jet engine limit the length of the
combustor. If thorough mixing is to occur, the order of turbu-
lence must be high. Therefore, not only must high-velocity
streams be usedt but they must also be injected In such a manner
as to "stir" the mixture thoroughly before it passes downstream
to the turbine. Insufficient mixing may allow stratification of
the layers of hot gases, possibly even resulting in a tongue of
flame impinging on the turbine blades.
Combustion efficiency will decrease, however, if the
secondary air is introduced so far upstream that chilling of the
products of combustion occurs before the chemical reactions have
been completed. Further, a high turbulence level results in a
large friction pressure drop through the combustion chamber.
Since friction pressure losses are normally of the order of twice
the momentum pressure losses due to heating,^ the friction pressure
losses greatly influence the efficiency of the gas-turbine cycle.
^Thus the design of a combustion chamber requires, among
other things, a delicate balance in the design of the secondary
air system. A compromise among burner lengtht combustion efficien-
cy, pressure losses, and combustion chamber exit (turbine inlet)
temperature profile must be made in the design of the secondary
air system. At this writing there is no data available to the
designer which will insure that if secondary air is admitted in
a prescribed pattern that an acceptable combustion chamber will
result. This paper will attempt to contribute some data for
this design problem, by determining the effect of a controlled,
reproducible turbulent air pattern on the factors mentioned in
the preceding paragraphs.
EQUIPMENT AND EXPERlKEaJTAL TECHNKiUE
EqjJIPMENT
The combustion chamber used in this experiment was designed
and constructed by Janssen." The chamber ig shown in Figs. 1, 2
and 14. Air was admitted through 48 ducts, the flow through which
was controllable by a damper plate in conjunction with a metering
orifice in each duct. The burner was rectangular in cross section,
being approximately 2 inches wide, 5 inches high and 20 inches
long. The front portion was a serai-circular arc, of radius 2.5
inches, as viewed from the side.
The main air supply consisted of a centrifugal compressor
driven by a 165 horsepower Lycoming air-cooled gasoline engine.
The air was ducted through a 6-inch pipe to a I-type manifold
where the flow was divided and routed to the lower and upper
halves of the burner.
The vaporizer tube for fuel injection was constructed of
5/8-inch outside diameter seamless stainless steel tubing as shown
in Tig. 5, The dimensions of the tube and its location in the
combustion chamber were determined by Miller and are shown in
Fig. 6.
The air through the vaporizer tube was obtained from a
hifijh-pressure (lOO psig) air line, throttled through a control
valve to a 2-inch pipe connected to that portion of the vaporizer
tube outside the combustion chamber. '^
The original eecondary air injection orifices in the com-
bustion chamber were modified by inserting steel stripe in the
original slots to establish a high turbulence level in the quench
air zone. This modification is sketched in Fig, 7 and shown in
Figs. 8 and 9.
The liquid fuels were pumped to the vaporizer tube by a
Vickers constant displacement pump. Fuel flow rate was controlled
by a hand-operated external by-pass system. A l/8-inch needle-
valve was installed in the fuel line in order to obtain steady fuel
flow at a pressure of 15 psig.
INSTEUMENTATICN
9 "7
The manometer system of Janssen and Byberg' was changed by
replacing the common water-mercury manometer system with one con-
taining a "separate water-filled U-tube for each of the 48 ducts and
the metering orifices of the main and vaporizer tube air supplies
(Fig. 3). The traversing tenperature and total pressure probes
in the exit section were replaced with a rake^*^ consisting of
seven total pressure tubes and seven chromel-alumel temperature
probes equally spaced in the vertical plane of the exit section
(See Figs. 10 and 11). Static pressure at the exit was obtained
by tapping two holes in the side plate of the exit section* one at
one-fourth of the distance down on one side and the other three-
fourths of the distance down on the opposite side, Joining the two
static pressure lines, and feeding the resultant pressure to
the water-filled manometer system. There the static pressure
was compared with atmospheric pressure and with the pressure
from each of the total pressure tubes of the rake.
The main air and vaporizer tube air flow rates were de-
termined by measuring the upstream static pressure and tenper-
ature and the. pressure drop in inches of water across a square-
edged circular orifice in the respective air supply lines. Mass
flow rates were computed in accordance with the procedure outlined
in Hef. 10 and are shown in I'igs. 12 and 13.
The fuel flow into the combustor was measured by a roto-
meter type fuel flow meter. The meter was calibrated previously •
and, in addition, thirty-minute runs at the design flow rate were
made for each fuel to confirm the accuracy of the calibration runs.
iChirty-nine chromel-alumel thermocouples were located in
three horizontal rows in the burner as shown in Figs, 14 and 23
and pictured in Figs. 2, 5, 8, and 9, The emf of the thermo-
couples was determined by potentiometers which had built-in cold
junction compensation for iron-iron constantin thermocouples,
Q
Thermocouple conversion curves shown in Fig. 15 were therefore
used to correct the observed temperature readings.
Iron-iron constantin thermocouples were used to measure
compressor outlet air ten^jerature and vaporizer tube air sujjply
ten^jerature upstream of the respective metering orifices.
EXPERIMENTAL TECHNI(.JJE
Since the primary purpose of this experiment was to de-
termine the effects of changing the method of introducing the
quench air, as many other variables as possible were kept constant
Thus, only in the configuration of the secondary air injection
orifices did the runs differ from those of Barnes. Combustion
intensity, fuel and air flow rates, vaporizer tube configuration
and location, and pressure drops in each of the 48 ducts adhered
as closely as possible to these values used by Barnes. In
addition, in order to use the same technique for each fuel the
runs on a particular fuel were always started with the upstream
quench air port at position 20 and the quench air %fas then moved
forward in increments of two stations per run. A period of ap-
proximately two minutes was allowed after all settings had been
made before any pressure or temperature readings were recorded.
This was done in order to allow steady conditions to obtain. All
temperature and pressure readings at the exit cross section were
read simultaneously and this cycle of readings was repeated at
least once each run. Teinperature readings of the 39 thermo-
couples in the burner were taken approximately every other run
on each fuel to determine the approximate boundary of the flame
pattern and the effectiveness of the secondary air as a quenching
medium. Fuel tenperature, barometer, and relative humidity
readings were taken before each set of runs.
10
The four fuels used in this projeot were aviation gasoline,
naphtha, kerosene, ajid a diesel fuel. Fuel specifications are
included as Appendix A and distillation curves as Fig. 16. The
design flow rate of each fuel was chosen to niatch that selected
ty Barnes, which in turn %fa8 selected so as to maintain the same
comhustion intensity (BTU heat release par second per cubic foot
of combustion chamber volume) for each fuel. Fuel flowmeter
calibration curves indicating these design flow rates are shown
in Fig. 17.
The actual operating procedure of setting up and running
g
was identical to that used by Barnes. Operating RPM of the main
air supply conpressor was necessarily increased approximately
300 BPM over that used by Barnes since the modification of the
secondary air injection orifices by the insertion of the metal
strips caused a flow restriction in the air supply system. In
order to maintain the same mass air flow through the burner it
was therefore necessary to operate the con5)res8or at a higher
pressure ratio, this higher pressure ratio being obtained by
increasing the conpressor BPM.
A series of runs consisted of burning one fuel at the
design flow rate, commencing with the first upstream quench air
port at air Station 20 (See Fig. 14) and moving the block of
quench air forward in increments of two stations per run.
The quench air pattern consisted of five ports on top
11
and the five opposing ports on the bottom of the burner, each such
duct having a pressure drop across the metering orifice equal to
six-Inches of vater. The remainder of the secondary air ports
were adjusted to 0.1 Inches of vrater pressure drop across each
metering orifice. The small flow rate through these latter ducts
provided air which served the dual purpose of tjreventing burning
In the ducts and keeping the ducts cleared of unvaporlzed fuel
during those runs where kerosene and diesel fuel were used.
The primary air pattern was that determined by Ryberg
and subsequently used by Barnes. All pressure-drop settings
were adjusted to give the desired values while combustion was
occuring. Typical air patterns are shown in Fig. 23.
Total air flow was maintained as constant as possible by
(l) hanff operation of the needle-valve controlling the air supnly
to the vjkporizer tube to maintain an air flow rate of 0.0218
pounds per second, and (2) varying compressor BPM to maintain a
main air supply flow rate of 0.600 pounds per second. These
values approximate the average air flow rates used by Barnes.
Inability to maintain these exact values is discussed in the
section on Brrors.
During the initial kerosene run it was noted that vapor-
ized fuel droplets emerged from cracks around thermocouple
insulators. The run was stopped and ceramic cement was applied
to the cracks. The next run caused the ceramic to crumble at
12
several locatione due to vibration, further atten^jta to reaeal
these cracks proved Just as unsuccessful. 'The kerosene runs were
therefore continued with fuel vapor leaking from the chamber.
With the quench air at station 20 very little fuel leakage was
apparent* but as the quench air was shifted forward the fuel
loss increased until at station 10 a cloud of vapor surrounded
the thermocouple side of the burner.
Four attempts were made to run diesel fuel at the design
flow rate, but all resulted in rich blow-outs. Combustion could
be maintained for only about 30 seconds after the butane (used
for starting) was turned off before flame-out occurred. Fuel
flow rate was reduced in small increments from the design rate
of 0.00517 lb/sec to 0.00221 lb/sec before combustion could be
maintained. This inability to bum at or near design flow rate
was appacently the result of the cementing of the cracks around
the thermocouples mentioned above. Since the amount of fuel
leakage was reduced considerably, the entrapped fuel vaporized
sufficiently to cause the rich blow-outs. As a result, the
runs on diesel fuel are not C0D5)arable with those of aviation
gas, naphtha, or kerosene, since the combustion intensity was
reduced to 43/^ of the design value.
13
ESTIMATED ERRORS
All pressure measurements could be read within £ 0.05
inches of v/ater or mercury. Ten^ieraturea could be read within 1°
of the scale value. It is estimated that fuel flowmeter readings
were accurate to within ^ 0.025 gallons per hour.
While the average fluctuation of the compressor was
l_ 5 RPM from the desired value, this produced no noticeable
variation in the main air supply readings. The velocity head
readings on the rake pitot tubes varied j^ OdO inches of water,
which would produce a maximum variation of two feet per second
in velocity determinations at high velocities and low densities.
The average rake temperature variation was £ 40p. between sue-
cessive readings. The main and secondary air metering orifices
were, standard orifice meters and should provide air flow readings
accurate to within j( 2^. Fuel flow rate did not fluctuate,
since any movement of the float was more in the nature of a small
vibration than an oscillation.
Considering only aviation gasoline, naphtha and kerosene
runs, the BTU input per pound of air was 155 j^ 10, and the overall
air/fuel ratio was 120.5 ^ 7.
Steep temperature and velocity gradients existed with the
quench air at stations 18 and 20. The mass flow integration and
resultant temperature rise computations are likely to be more in
14
error at statione 18 and 20 than they are with the quench air at
stations 16 and forward.
Efficiencies greater than 100^ were encountered on aereral
runs. Two factors contribute to this error:
(l) No horizontal traverses of the exit section were
attempted. Thus the measured temperatures at the
center of the section were considered to extend
the width of the cross-section in the exit-section
summation o^f'i^i'^i* Hyherg on this same equipment
made one horizontal traverse per run, and determined
that the temperature decreases from the centerline
toward the edges. This error leads to efficiences
that are too high in every case.
i2) No high- temperature calihration of the exit section
*^ thermocouples was made. The rake was checked in
hoiling water; variation of Indicated temperatures
on all thermocouples did not exceed 1.50F. from the
average temperature. Since the butt-welded thermo-
couples were surrounded by ceramic shields, it would
have destroyed the thermocouples to have checked
them in a molten metal.
The moisture content of the air remained at 51 ^^ 8 grains
of water per pound of dry air. Since this variation was small,
omitting the effect of moisture in efficiency computations lead
15
to a constant error of approxime.tely 0.7^ in the efficiency.
This was considered negligible. All computations were therefore
based on fuel and dry air as inputs to the combustion chamber.
15
RESULTS
The purpose of thie investigation was to determine (l) a
method of introducing secondary air in a combustion chamber in
such a manner as to cause definite quenching and (3} the effects
of such an air pattern on combustion efficiencyi combustor length
and combustor exit temperature profile, and combustion chamber
pressure losses. Since the method used in this experiment was
to be compared with that of a preceding method on the same equip-
ment, it was necessary to adopt the same definitions of "efficency"
and "optimum combustor length" as that used by Barnes.
Thermal efficiency as used In this paper is therefore
defined as the ratio of the actual temperature rise to the theo-
retical maximum temperature rise of a given mass of air burning
in the combustor with a piven mass of fuel. The method of de-
termining thermal efficiency is included as Appendix B. Optimum
combustor length is that distance from the most forward point of
the combustion chamber downstream to the centerline of the first
quench air port at which the thermal efficiency of the cycle is
at or very near a maximum.
Plots of thermal efficiency vs. coiAustor length are
shown in Tigs. 18, 19, 20, 21, and 22 for the four fuels.
Aviation gasoline, of lower molecular weight and higher vola-
tility than naphtha, indicates a slightly higher thermal
17
efficiency (102.055) as compared with naphtha (101.55^). Both of
these fuels exhibit the same shape curve; that is, constant or
slightly increasing efficiency from station 20 to station 12, at
which point a rapid fall-off in efficiency is apparent. The exit
temperature profiles also are similar, with sharp gradients oc-
curring with the quench air at stations 20 and 18, gradients
decreasing as the quench air is moved forward, until an almost
even temperature profile is obtained with the quench air at
station 14 and forward.
The curve of r> vs. corahustor length for kerosene indicates
a linearly-decreasing efficiency from station 20 forward. This is
considered to "be due to quenching and to the loss of fuel from
around the thermocouples mentioned previously.
From the plots of thermal efficiency vs. comhustor length,
the optimam comhustor length for the various fuels was as follows!
J'u?;
Air Stftt^pn
Aviation gasoline
12
Naphtha
12
Kerosene
20
The outputs of the 39 thermocouples inside the combustion
chamber were recorded with the quench air at stations 20, 16, 12
and 10 for each of the four fuels. It was assumed that the
minimum temperature for combustion to exist was 1850°R. The
18
flame patterns for theae runs was then estimated, yiame patterns
for runs on naphtha are shown in Fig. 23. For the three comparable
fuels it was shown that the quench air definitely "chopped off" the
flame as it passed between the 3rd and 4th quench air ports. The
diesel fuel flame pattern was generally shorter and not as well
defined as those of the other fuels.
With regard to oomhustor exit (turbine inlet) allowable
temperature (assumed to be 1700°F. maximum), the temperature
readings in the chamber indicated that the combustion chamber
could have been physically shortened as indicated below without
exceeding the maxinoum temperature limitation at the exit:
First Quench Air Port
Naphtha - cut at thermocouple #37 12
"Aviation gas " " #37 12
Kerosene " " #06 20
This "shortening" would have been based on the quench air being
positioned in accordance with the optimum combustor length as
defined previously. The limit of 5f> variation from average
temperature would not, however, have been satisfied.
The total pressure loss on each run was computed in
accordance with the procedure outlined in Appendix C and is
plotted in Fig. 24. Points were selected from the raw data
of Barnes which were near or on each of his final curves of
19
efficiency vs. combuator length and the pressure loss was computed
in the same manner and is also plotted in Pig. 24 for purposes of
comparison. It can tie seen that introducing the strips in the
injection orifices increased the total pressure loss from an aver-
age value of approximately 5.6^ to an average value of approximately
11. 3^^. The momentum pressure losses were very nearly equali within
the limits of accuracy of both investigations (See Appendix C) .
Therefore, the change in total pressure loss is equal to the
increased friction pressure loss due to the modification of the
injection orifices and the resultant turbulence in the burner.
Summarizing! the results of this investigation clearly
showed that with either aviation gasoline or naphtha as a fuel,
quenching was positively effected and completely extinguished the
flame between the third and fourth quench air ports. Relatively
high efflpiences were also attained. The temperature and velocity
profiles at the exit of the chamber were acceptable as turbine
inlet profiles. Total pressure losses were doubled (increased
from 5.6^ to 11.3/^) due to the highly turbulent flow.
The vaporizer tube was not capable of efficiently handling
kerosene as a fuel, and would not even operate with diesel fuel at
the design flow rate.
20
BIBLIOGRAPHT
1. Oodaey, F. W. , and Young, L. A., Gaa Tm-binps for Air^rAft.
New Torlc: McGraw-Hill Book Co., Inc., 1949.
2. Olson, L. 0., F. W. Ruegg, and F. R. Caldwell. "Combustion
in Moving Air," S.A.fi. Qimrtfirly Tranflactlnna. April, 1949.
3. Clarke, J. S. , "Combustion in Aero Gas Turbines," Engineering
September 15, 1950, 170:230-232.
4. Elliott, M. A., "Combustion of Diesel Fuel," S.A.E. QiiarterLv
Transactions. Vol. 3, 1949.
5. Miller, W. L. , Master of Science Thesis submitted to the
University of Minnesota, 1954.
6. Barnes, R. J., Master of Science Thesis submitted to the
University of Minnesota, 1954.
7. Ryberg, J. G. , Master of Science Thesis submitted to the
University of Minnesota, 1953.
8. Vincent. £. T. , Tha Th*tnry anrt Design nf Gaa TtirbinflB and
Jftt gng-lnew. New Tork: McGraw-Hill Book Co., Inc., 1950.
9. Jansseli. J. E. , Master of Science Thesis submitted to the
University of Minnesota, 1953.
10. Flow Mftflsurement 1949. A.S.M.E. Power Tgst Codes.
11. Pinkel, I. V., and H. Shames, Analyals of Jet-Prnpnlaion
Enflnft CnmbnRt.tnn-Chinnbpr PreHRiir» LoHBeH. NACA TK 1180,
Government Printing Office. Washington, D. C, February, 1947.
12. Worth Amfirinan CnmbuBtlnn Bantlbooki The North American
Manufacturing Co. , Cleveland, Ohio, 1952.
13. Griswold, John, FuwlHr Combnati nn and FurnAceg. New York:
McGraw-Hill Book Co., Inc., 194€o
14. ASTM Test Code: D86-40.
15. ASTM Test Code: D323-43.
22
(a) Sidt View of Combustion Chambtr.
(b) Combustion Chambtr With Sid« Plat* Removed,
Figure 2
23
Figure 3 - T«»t Cell Control Panel,
Figure 4 - Manometer Syetem.
24
r
(a) Vaporizer Tub*.
INI I 'v"'v:i
/
But an*
Inl«t
(b) Vaporizer Tubt Installed
Figure 5
25
(^
o
Co
::!
O
oa
K
5:
^
o
o
o
o
*^
o
s
:^
Q:
o
:^
h
5
Q
^
^J
^
kj
UJ
k.
^
c^
s
-^
<>7
^
c»>
Mj
^
.•
^
cb
o
5:
^
^
27
Figure B - Combustion Chamber Showing Lower Injection
Orifices as Modified.
Figure 9 - Close-up of Modified Lower Injection Orifices,
28
Figure 10 - Exit Cross-section Rak«, Assembled.
Figure 11 - Exit Cross-section Rake, Showing Construction,
31
^1
^
^
cM^. ^^ 'n-|^
^-1-
— o
(0-|-
T""
c^f ^f *^f
-f-
4- nf
•1^ tnJ.
S+
r-4. -n-^ ^f
^~!:'^
"i-^+J; "+-
<M
+ S+-- -+
91
5
0)
I
I
O
I
%
«^4
2/'^;?^^54-
40
SrATio^t No.
^
y
'^si-i-i^;^3.^^^^^4w^
5 J f 'f •• f • •» •• •* •• ••
• ' >-| I I I i t > I > I— L
/ J ./ j</ Dnop,
i. I I — L
I
I I I I I I I I I — r-T
r I I — r
. I i I . .
t i I i uu.
Z^ 3- V-^-^i' 1^ a- 10 2\ 2i 23 i*
' /jf /8 /5>* 34^ 36 j^ 37' BB ^ ZJ 24 2^
V I ' I • ' M I I
. . \ COUPLE No.
I 1 I I I I I I
Cme Position,
-T — I I r 1 < I I I I
FiCURE 23, FLA^e. PaTTB RNS ] F^eL ^NAPHTHh
43
APPENDIX A
FUEL SPECIFICATI0NS6
The four fuels used in this inyestigation are the same as
those used in the report by J. Ryberg, Although the values of
the specific gravity of a particular fuel as used in each investi-
gation differed somewhat, the overall heating value and specific
heats did not vary enough to introduce errors for purposes of
comparison.
The heating value, latent heat of vaporization, and the
weight per cent of hydrogen and carbon were computed from an
equation using the specific gravity. ^^ As these values did not
vary anpreciably from the various handbook values, they were
accepted as being sufficiently accurate for purposes of calcu-
lations in this investigation. The equations used are:
(1) net heating value in BTU/lb = 19960 --3,780 x (spgr)^
- 1362 X (spgr). (for const, press.)
(2) latent heat of vaporization in BTU/lb
= 110.9 - 0.09 Y teigp.^F where t°F was chosen as
sp gr an average between the
boiling point and the
terap. of the incoming
air.
(3) weight ^ of Hydrogen = 26 - (l5 x spgr)
(4) weight $ of carbon = 100 - wt. ?^ of hydrogen
The specific heats of the fuels were based on equations
13
wherein a factor K was used. This factor is a direct indication
44
of the fuel characteristics as shown by the equation
K - ^ Tw j^^^ 0£ It is plotted as a parameter in curves of
sp gr at 60°F.
temperature versus c . From the equation
'Pav, ■ ^/« Sti ^Pt„ ^ Sta'
the average Cp may be calculated. The upper and lower values of
the temperatures of the fuel vapor were taken as 800°?. and 1500°F.
(^T^. = ^T^ ^rom curve at ti / tp while C_ = final GL^ ).
Ptg^y P -^^-^ — ^ Pave i^fuel
The distillation curves shown in Figure 16 were determined
according to the specification given by the ASTM Distillation Code.
The vapor pressures of the fuels were determined with a
Reid vapor pressure bomb in accordance with the ASTM specifications
and standards. ^^
Densities were determined by use of the Westphal balance.
Th« fuel properties are found in Table I.
45
o
00
o
l-l
P-
c^
lO
in
lO
c^
o
C\}
00
o
03
o
C\3
to
a>
a>
o>
crt
O
lO
•
•
•
•
•*
CM
•
o
to
C^
o
CO
f-t
CM
00
d
c
e
a
a
o
u
CM
{>•
CM
00
o
CJi
CM
o
00
U3
OD
to
to
o
O
•
•
•*
iH
in
<o
o
00
CO
00
in
CM
■^
00
i£>
t^
CM
^
ID
O
in
in
(T>
C^
to
to
to
in
CM
o
fH
00
•*
•
•
«
•
CM
iH
00
•
•
•
o
s
to
o
00
r-i
in
1-4
CD
O
«
^c§
>
<J>
<*
CM
o
CM
to
in
^
%D
CM
to
r-
CM
r-
c-
O
r-(
o
CO
lO
rH
00
in
•
•
•
CM
r-t
CO
•
•
•
o
to
«3
to
U3
00
13
00
o
^
g
m
•
,,,^
Pc«
•
o
1
a
a
«>
ft
t
•
•
a
■»>
^
ttO
o
03
.
o
a
•H
c
=♦:
•
^
o
o
CO
•
o
'
(^
o
«D
iH
ft
43
u
D
o
o
r-l
CO
>..•
EH
(O
*»
S
1
1-1
pq
=*!
(d
ti
tl
^■^^
■a
<D
o
1
p
00
•
»4
U
a
ft
EH
>
«
d
o
cd
0)
«
>»
(C
ft
n
u
>
=<
c
4*
•H
^
rH
^
<M
rH
id
&
c3
1
>•
<D
(4
+>
O
>
o
o
■♦*
g
^
P.
•H
»
evD
^
f,
CO
dC
Vh
»4
o
2
c
b
03
Si
•H
<H
o
ft
•H
jd
o
O
O
O
ft
si
+>
o
•H
0)
t
w
2
<M
«^«
■H
V<
Pi
(D
d
■*>
o
o
<H
•H
«
tJ
fH
c
^
f-«
O
•d
cH
«
VP.
"^
o
(D
»-«
•H
•H
•**
*3
<D
ft
23
O
P4
^
O
^
0)
4*
4»
ft
CO
46
APPENDIX B
iiiFFIGIBlJCY CALCULATION'S
Calculation of thermal efficiency, defined as
;: AT ac!tu/>X , for each run was carried out as follows:
^ AT
theoretical
Data recorded:
Fuel, Fuel Flowmeter reading. Air Pattern.
Data computed:
i-2./3.Jl. ^ltf"-^i/ Vl'^^Il^-
Ifeta obtained from calibration curves:
• • •
n>2i ififj* nijr*
Computation procedure:
Average exit cross-section temperature:
^^ * = aveo
Actual temperature rise in "burner:
■fact = fave " <f2Vp ^ ^pV-^) = V« " (^2 ' l"'
<^S^2 / J s's)
Total Mass flow at exit cross section:
mj3 =T^^V^k^ =L\ i^i (I)(2.?5 I 5) = 0.01116^^-^7^, ]Jb
^ 7 144 sec.
Air mass flow at exit cross section:
"^ir = n'm- "'f » ii
sec.
47
C determined from Fig. 25.
^air
Pmixture " ^ Pf ^ air P^jr "J^T^r.
.71
m
Theoretical temperature rise:
^theoretical = ^f'^L . ^B..
" mixture
, = AT„,,„., X 100. ^.
^ AT
theoretical
%.
46
APPENDIX C
PHBSSUiUS LOSS COIvlPUTATIUlJS
Data Recorded:
Pg.APg. T^, P3.AP3. Tg. Pfi • ^1' (^P)i' and T^,
as defined In SYKBOLS.
liata coii5)uted or taken from calibration carves!
J 2* y 3* °^' "^2 ^"^^ 1 i ^® defined in SYMBOLS.
Assun^ption: Velocity profiles upstream of metering orifices
are constant.
Formulas used:
(1) m - f VA, lb/sec.
(2) Pm = P^ _ = P ^ • / I V V2.
•'' total static ^
^^^ ^total =^^total/j =^^totalj ^Vj
Sample calculation (Run #3):
2x . "" , -, /. 2
_ m„ - ^Po it * f oV,
3 L )2-2lj
^T^ ^ = (^2 / i j 2^2 ) ^ = ^2 '^ ^ j 2 ^"^ h =
- V*2 '^ V
^i2^2'^
r (.5910)(32.55) / (.5910)^ (2S.?a)
2(. 0750) (32. 2) (2116) (.1962)^
- 19.378
• 3 , __,3/
3 3 3 ^ 2 2(.0750)(32.2)(2116)(.1962)'
Pm m, ; i P / m„ - (.nP14)^'^(?9.g2)
/ (.0214) (32. 17)
- 0.688 / 0.00000224 - 0.688.
49
Inlet total pressure - Pm mrj / Pm m
^2 i ^3
Pm = 19.. -^78 4. n.fiflfl = 32.80" Hg.
hNLBT .6124
Exit Total Pressure:
= ^U'^l^^aiiiblent '' ^''static "^ ^^Impsctt^
- ^ambient '^'^-^ i^i ^^^static ^ ^^impact^^
^ji^l
- 28.87 / 8.754
52.48
z 28.87 / 0.167 = 29.04" Hg.
Total Pressure Loss:
Pm loss - Pn( - Pt
^ Hnlet STfU
Pj
inlet
- .-^P.aO - ^9.04 X 100 - 11.48^
32.80
From Ref . 1» the equation for momentum pressure loss in
a combustion chamber of variable cross section is
2 ' Ihh J
where M^ - inlet Mach number
P3 - inlet total pressure
Ag - inlet cross section area
Tt - inlet cross section tenperature
Oa r refers to exit cross-section
50
To determine the increase in friction pressure loss
between those runs by Barnes and those of this investigation the
following procedure was used. The total pressure loss for each
run in this investigation and for each run which plotted on or
near final curves of Barnes was confuted in accordance with the
procedure previously outlined. Data from runs at hi^h efficiencies
(large T^) are herewith compared. Prom Hun #4 of this investigation,
using only main air supply data:
T„ - 577OR. Big - .5910 lb/sec.
T4 _ 12410R. L^ ^ .0750 Ib/ft^
i?=ir25 = ^-^^^ ^^loss = ''-'"^
M Z 7 Jl2_ - 0.0350
as 4 '^
J^ = kC.O.-^SO)^ 1241 X 2.510 - 1
Pg 2 L 577 J
s - k(. 000612) 5.40-1 = 0.00257 k
- L J
From Run #47 of Barnes:
T - 546°R. m„ = 0.5980 lb/sec.
T^ I II67OR. 3 = 0.0749 Ib/ft^
if = 2.510 ^ 5,51^
4 loss
M_ = V„ z m„ - 0.0354
•-^ i2 17 ■ ~
^3 lis
^=^ T^A^-1
^3 2 1T3 A4 ^
- k(.0354)^ 1167 X 2.510 - 1 = .00626(4. 35)k
,546 J
z 0.00272 k.
51
% change in momentum pressure lose* ^ased on that of
Barnes is:
^ change = n.nnP7P k - 0.00257 k X 100 = 5.52^ (decrease)
0.00272 k
To determine the momentum pressure loss, since the factor
"k" is unknown, the relationship that friction pressure loss is of
the order of twice th
for Barnes'^ Run #47:
the order of twice the momentum pressure loss is used, yielding.
% Momentum pressure loss = 3 (5.51)^ :; 1.83^
Applying the 5.52^ decrease computed above, the momentum pressure
loss for Bun #4 is 1.735^. This difference of 0.10^ (= 1.83^ - 1.73^)
is insignificant when compared to the total pressure loss (= 11.70^).
Therefore, the increase of friction pressure losses between the two
investigations has been taken as the difference in the total
pressure losses.
J
CP 29
DISPLAY
S5006
Hutches
An investigation of thel
effects of turbulent
ouenching in s cpn-type
combustion chpmber.
EP 29 DISPLAY
Thesis
H958
Hutches
An inveptifption o"^ the e'*''^«"'t'
of turbulent oieaching in a cen-
type combustion chanber.